PY Wall Deflections

[DirtDr]

We recently started using PY wall, but before we started using it on a regular basis, we compared it to some inclinometer data we had in our files. I also compared it to some PLAXIS simulations on a current project. I noticed that the deflected shape of the soldier piles using PY Wall was close to the measured shape, but the total deflections were off by a large margin. In fact, I had to double the lateral load on the wall to get the same deflections as measured by the inclinometer or estimated by PLAXIS. The walls were all tied back, top-down, soldier pile and lagging walls. I used the FHWA manual to estimate the lateral loads. I tried to "soften" the soils, but the soil parameters became unreasonably small, and we had several UU tests and DMT’s to develop the soils parameters. Any ideas as to why PY wall would be off so much?

 

[PEinc]

Pressure diagrams for tiedback walls are not true representations of the actual wall pressures. The diagrams were intended to give the maximum anticipated brace or tieback load, not the actual load at every brace. The diagrams often are determined by taking the thoeretical triangular load per LF of wall, multiplying it by at least 120%, and then reshaping the load into a rectangular or trapezoidal form. Therefore, the earth pressure diagrams are conservative. Inclinometer deflections are the real deflections. I never used Plaxis, but if the Plaxis deflections match the inclinometer deflections, then Plaxis must be more accurate than PY Wall. Or, the specific soil properties in Plaxis were more accurate than those used in PY Wall. In my experience, people who specialize in the design and building of shoring walls usually do not use either PY Wall or Plaxis.

 

[DRC1]

Could be several reasons
1.) despite testing, the tests may not accurately reflect soil conditions. always a challenge in geo tech.
2.) the designer was conservative in the parameters he chose - not uncommon for reasons in no.1
3.) The wall may not have devloped sufficent strain to load the walls. We have all seen walls that but for the grace of God, should have fallen. I have seen deep trenches up to 12 feet in clean fine sand stay vertical for extended periods of time.
What you are designing for is when there is sufficent strain to fully load the piles
I am not sure where FHWA got their new wall load profiles, but they a re a significant departure from previous assumptions.
Seeing how you have the right profile, but it is half the anticipated deflection, I suggest the designer may have been conservative.

 

[SmokeyBear]

PEinc, Agreed - apparent pressure diagrams are conservative envelopes, widely used before PCs. Soil-structure interaction analyses must be carried out for multi-propped walls and should be used for single props. You commented "..people who specialize in the design and building of shoring walls usually do not use either PY Wall or Plaxis." What is commonly used? We use WALLAP, GGU sometimes and PLAXIS for serious stuff.

 

[PEinc]

Most often I see the use of CT Shoring, ProSheet, SPW911, and hand or spreadsheet calculations. Other methods are usually very expensive and more complicated than necessary compared to the usually guestimated soil values. Rarely do I see any projects where unit weights, friction angles, and cohesions are actual test values. I know it sounds wrong, foolish, and/or risky, but most sheeting walls are designed with soil values estimated from boring logs, tempered with local experience. In my experience(designed and/or built well over one thousand walls), seldom do the soil values contribute to wall performance problems. Usually problems are caused by poor workmanship in the field, by greatly differing ground conditions, by unexpected surcharge loads, or by use of an inappropriate earth pressure distribution.

 

[Doc08]

My two bits on the matter.

We do a fair amount of shoring works and we have used several of the commercially available software including PYwall and Plaxis. The results: We ended up writing our own software. Why?? none is complete, none addresses all of the practical day to day needs. Perhaps they should not be expected to be so ????

We have collected several published case studies, some of which have been already analysed by the authors using Plaxis or similar FE software. I am currently convinced that P-Y method (for most cases) is just as good as any and the impact of a good soil investigation and determining the proper design soil parameters far outweigh anything else. We also found out that, in general, it is very difficult to match both: the reactions and the deflections of the retaining system, that is when compared with the actual field cases. Again, as an overall conclusion, when you match the deflections, the calculated reactions are on the conservative side by, anywhere from 10% and up.

Furthermore, and especially when braced excavations are involved, it seems that there is more room for construction misshaps ranging from delays in the execution to faukty excutions to fauilure in accounting for sagging in the long braces which inevitably translates into a lateral translation in the shoring system, etc...

I'll be glad to provide a copy of this exercise/document should anyone be interested.

Moral of the story.

1. There is no substitute for a good soil investigation, good testing program, experience (especially local experience), and common sense.

2. More expensive software is not necessarily better nor should it be treated as the bible. The engineer should have the knowledge to analyze and determine what makes sense.

3. Know your software and its limitations and be sure you can verify its results, even if roughly, using quick hand calculations.


Good luck

 

[DirtDr]

Doc08,

Thank you very much for your insightful comments. I couldn't agree more that these tools are not gospel and should be considered little more than guides to aid our judgement. I use limit equilibrium techniques as the methods of choice and use these other, fancy computer methods as second and third opinions. I started to explore these two applications to help develop questions to be answered - typically by instrumentation in the field during construction as well as a design aid for walls.

I am very interested in reading the material you have assembled. Please attach it or send a link.

 

[Doc08]

http://www.tsoftonline.com/Software List.htm

and click on (download Case Sudies & Validation Document)

good luck